3DiVi launches image quality assessment algorithm for face biometrics

3DiVi Inc. is introducing a Quality Assessment Algorithm (QAA) to help reduce errors by face biometric authentication systems.

Banks and fintechs often work with datasets that contain facial images impacted by low lighting, extreme angles or occlusions like masks or glasses that increase the risk of false positives, false negatives and fraud, according to the announcement. Evaluating the quality of images before biometric comparisons helps mitigate these risks.

The new QAA from 3DiVi automatically filters photos for suitability with facial recognition, and can cut facial comparison errors by 50 percent and liveness detection errors by 40 percent, the company says.

Computer vision and facial recognition experts from 3DiVi recommend comprehensive analysis of all images and removal or replacement of all poor-quality images to improve the quality of biometric datasets. Organizations should also implement ongoing validation for continuous quality control and provide guidelines to end-users for environmental adjustments like optimal lighting and camera positioning.

ISO/IEC 29794-5 is the biometric sample quality standard for face image data.

NIST evaluates face image quality assessment algorithms as part of its Face Analysis Technology Evaluation (FATE) program, and published a new report on Specific Image Defect Detection (SIDD) this month.

Article Topics

3DiVi  |  biometric authentication  |  biometric data quality  |  biometric liveness detection  |  biometrics  |  face biometrics  |  ISO 29794-5